Please help: Archive DAT crashes a system

Thread View: comp.unix.sys5.r4

2 messages

[T] Back to Threads [V] Tree View [G] Group Articles

2 total messages Started by Evan Leibovitch Tue, 15 Mar 1994 17:16

#837

Author: Evan Leibovitch
Date: Tue, 15 Mar 1994 17:16

31 lines
2097 bytes

Hardware:
Everex 486/33 host (EISA)
Equinox Megaplex serial ports
Adaptec 1742 EISA SCSI controller
Micropolis internal HD
Archive external DAT drive

Software:
Esix r4.0.4

Sometimes, not every time, the system gets in a state where writing to
the DAT drive causes a kernel panic. Reading never causes a problem.

The problem is *not* repeatable on demand. Sometimes a tape write will
succeed, sometimes the identical command will crash.

Various combinations of SCSI load, reset, etc block-size-setting commands
have been tested in advance of the tape operation, with no discernable
difference.

Strange aside: The SCSI terminator must be left off the external
drive; if it is installed the system will not boot. The terminating
resistors have been removed from the 1742, and are installed on the
internal HD.

Any suggestions are welcome.

-- 
 Evan Leibovitch, Sound Software Ltd., located in beautiful Brampton, Ontario
         evan@telly.on.ca / uunet!utzoo!telly!evan / (905) 452-0504
"We're just not good committee people. Some of us don't have that long to live"

Re: Please help: Archive DAT crashes a system

#838

Author: Raymond Shwake
Date: Wed, 16 Mar 1994 02:56

23 lines
2212 bytes

evan@telly.on.ca (Evan Leibovitch) writes:

>Sometimes, not every time, the system gets in a state where writing to
>the DAT drive causes a kernel panic. Reading never causes a problem.

>The problem is *not* repeatable on demand. Sometimes a tape write will
>succeed, sometimes the identical command will crash.

	This reminds me of an ongoing problem we'd encountered on our
Austin 486/50 EISA tower. We were (still are) running Dell UNIX on that
box, with an Adaptec 1742A (in standard mode) supporting an HP R-Dat.
Almost consistently, when we ran heavy I/O (especially doing backups)
the system would crash.

	We were advised to experiment with internal and external caches.
Once we turned off external (128kb) cache, the problem disappeared. (L1
cache is still enabled.) Adaptec suggested the motherboard may not be
properly engineered to handle heavy DMA with the adapter driver in use.
(Our USL SVr4.2 box with an identical setup however doesn't demonstrate
the above problem.)
-- 

uunet!media!irscscm!nearside!shwake			shwake@rsxtech

Thread Navigation

This is a paginated view of messages in the thread with full content displayed inline.

Messages are displayed in chronological order, with the original post highlighted in green.

Use pagination controls to navigate through all messages in large threads.

Back to All Threads